What are the hardware requirements for Base?

Base requires approximately 49.4 GB of VRAM and supports a context window of 1000K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Base and how large is it?

Base is developed by Aleteian, a model with 12b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Base?

Base is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Base by Aleteian — VRAM 49.4GB, 1000K context

Name: Base
Author: Aleteian

Base is an open-source language model by Aleteian. Features: 12b LLM, VRAM: 49.4GB, Context: 1000K, Instruction-Based, Merged, LLM Explorer Score: 0.16.

Merged Model Autotrain compatible Base model:ilyagusev/saiga nem... Base model:ilyagusev/vikhr nem... Base model:natong19/mistral-ne... Base model:pocketdoc/dans-pers... Base model:vikhrmodels/vikhr-n... Conversational Endpoints compatible Instruct Mistral Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/Aleteian/base

Base Parameters and Internals

LLM Name	Base
Repository 🤗	https://huggingface.co/Aleteian/base
Base Model(s)	...Nemo Instruct 2407 Abliterated Saiga Nemo 12b PocketDoc/Dans-PersonalityEngine-V1.1.0-12b Vikhr Nemo Orpo Dostoevsky 12b ...r Nemo 12B Instruct R 21 09 24 natong19/Mistral-Nemo-Instruct-2407-abliterated IlyaGusev/saiga_nemo_12b PocketDoc/Dans-PersonalityEngine-V1.1.0-12b IlyaGusev/vikhr_nemo_orpo_dostoevsky_12b Vikhrmodels/Vikhr-Nemo-12B-Instruct-R-21-09-24
Merged Model	Yes
Model Size	12b
Required VRAM	49.4 GB
Updated	2024-12-30
Maintainer	Aleteian
Model Type	mistral
Instruction-Based	Yes
Model Files	2.7 GB: 1-of-11 4.9 GB: 2-of-11 5.0 GB: 3-of-11 4.9 GB: 4-of-11 5.0 GB: 5-of-11 4.9 GB: 6-of-11 5.0 GB: 7-of-11 4.9 GB: 8-of-11 5.0 GB: 9-of-11 4.9 GB: 10-of-11 2.2 GB: 11-of-11
Model Architecture	MistralForCausalLM
Context Length	1024000
Model Max Length	1024000
Transformers Version	4.46.2
Tokenizer Class	PreTrainedTokenizerFast
Padding Token	<pad>
Vocabulary Size	131072
Torch Data Type	float32

Rank the Base capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Base

Best Alternatives	Context / RAM	Downloads	Likes
...r Nemo 12B Instruct R 21 09 24	1000K / 24.5 GB	396277	141
Qliphoth 12B V1.1	1000K / 24.5 GB	22	3
Qliphoth 12B V1	1000K / 24.5 GB	26	2
SauerkrautLM Nemo 12B Instruct	1000K / 24.5 GB	1473	22
Magnum V4 12B	1000K / 24.5 GB	1116	54
MN Slush	1000K / 24.5 GB	181	32
...g M Claude Opus High Reasoning	1000K / 24.5 GB	82	3
ChatWaifu V1.4	1000K / 24.5 GB	335	20
ChatWaifu 12B V2.0	1000K / 24.5 GB	175	22
Mistral Nemo Bophades 12B	1000K / 24.5 GB	18	9

Note: green Score (e.g. "73.2") means that the model is better than Aleteian/base.